In the field of the manufacture of various electron devices typified by semiconductor devices in which fine working of half-micron order is required, the further increase of density and integration of these devices has been demanded. Therefore, a photolithography technique required for the formation of fine patterns has been more and more strict.
In particular, in the manufacture of DRAMs having an integration degree of 1 gigabit or more in which a working technique of 0.18 μm or less is required, the utilization of photolithography in which an ArF excimer laser (193 nm) is used has been investigated in recent years [Donald C. Hofer et al., “Journal of Photopolymer Science and Technology”. Vol. 9, No. 3, p. 367-397 (1996)]. Accordingly, it has been desired to develop a resist material suitable for the photolithography in which the ArF light is used.
In developing this resist for ArF exposure, it is necessary to improve a cost performance of the laser, because a life span of a gas which is a raw material of the laser is short and a laser apparatus itself is expensive. Therefore, high resolution properties suitable for the fine working and the enhancement of sensitivity are strongly desired.
As the resists having a heightened sensitivity, there are well known chemically amplified resists in which a photo-acid generator as a photosensitive material is utilized. For example, as typical examples, resists comprising combinations of triphenylsulfonium hexafluoroarcenate and poly(p-tert-butoxycarbonyloxy-α-methylstyrene) are mentioned in Japanese Patent Application Publication No. 27660/1990. Such chemically amplified resists are now widely used as resists for a KrF excimer laser [e.g., Hiroshi Ito, C. Grant Wilson, “American Chemical Society Symposium Series”, Vol. 242, p. 11-23 (1984)]. The chemically amplified resists are characterized in that a proton acid generated from the photo-acid generator as a contained component by light irradiation gives rise to an acid catalytic reaction with a resist resin or the like by a heat treatment after exposure. As understood from the above, in the chemically amplified resist, there can be achieved a much higher sensitivity as compared with a conventional resist having a photoreactive efficiency (a reaction per photon) of less than 1. Nowadays, most of the developed resists are of the chemical amplification type.
However, in the case of the lithography in which a short wavelength light of 220 nm or less typified by en ArF excimer laser is used, the resists for forming fine patterns are required to possess novel characteristics which the conventional material cannot satisfy, i.e., a high transparency to an exposure light of 220 nm or less, and a dry etching resistance.
A conventional photoresist material for g-line (438 nm), 1-line (365 nm) or the KrF excimer laser (248 nm) utilizes a resin such as a novolak rosin or a poly(p-vinylphenol) in which an aromatic ring is present in a structural unit. The dry etching resistance of this aromatic ring enables the etching resistance of the resin to be maintained. As a negative photoresist material, a crosslinking agent is further added to the resin. Examples of the usable crosslinking agent include azide compounds such as 2,6-di(4′-azidobenzal)-4-methylcyclohexanone and 3,3′-diazidodiphenylsulfone as well as methylolmelamine resins. However, the resin having the aromatic ring extremely strongly absorbs a light having a wavelength of 220 nm or less. Therefore, most of the exposure light is absorbed on the surface of the resist, and so the exposure light cannot reach a substrate, with the result that the fine resist pattern cannot be formed. For this reason, the conventional resin cannot be directly applied to the photolithography in which a short wavelength light of 220 nm or less is used. Accordingly, a photoresist material is now desired which contains no aromatic ring, has the etching resistance, and is transparent to the wavelength light of 220 nm or less.
As polymeric compounds having the transparency to the ArF excimer laser (193 nm) and the dry etching resistance, there have been suggested copolymers each having an adamantyl methacrylate unit which are alicyclic polymers [Takechi et al., “Journal of Photopolymer Science and Technology”, Vol. 5, No. 3, p. 439-446 (1992)] and copolymers each having an isobornyl methacrylate unit [R. D. Allen at al., “Journal of Photopolymer Science and Technology”, Vol. 8, No. 4, p. 623-636 (1995) and Vol. 9, No. 3, p. 465-474 (1996)].
However, the (meth)acrylate derivative having an alicyclic group which is used in the former resin does not have any polar group having adhesive properties to a substrate (e.g., a carboxyl group or a hydroxyl group). Therefore, a homopolymer of a monomer having the alicyclic group is strongly hydrophobic and poor in the adhesive properties to the substrate to be worked (e.g., a silicon substrate), and so it is difficult to form a homogeneous coating film with a high reproducibility. Furthermore, the former resin does not have a residue capable of expressing a solubility difference before and after exposure in an adamantane-including residue, an isobornyl-including residue or a menthyl-including residue unit having the dry etching resistance, and therefore, any pattern cannot be formed by the exposure. Thus, the former resin can be utilized as the resin component of the positive resist only by forming a copolymer of the former resin itself with a comonomer such as t-butyl methacrylate or a tetrahydropyranyl-methacrylate capable of exerting the solubility difference, or a comonomer such as methacrylic acid having the adhesive properties to the substrate. However, a content of the comonomer is required to be about 50 mol %, and the dry etching resistance per comonomer is noticeably low, so that a dry etching resistance effect by the aliphatic group noticeably deteriorates. Accordingly, the former resin is less practical as the resin having the dry etching resistance.